Description of Mr. Kemp’s Two New Thermometers. 263 
Fahrenheit’s scale between zero and the boiling points of liquids, 
its length would render it quite unfit for use. 
When a thermometer is to be employed to mark any slight dif- 
ferences which might take place in the boiling point of liquids un- 
der a variety of circumstances, as for example, the difference that 
exists between the boiling point of water in metallic and glass ves- 
sels, the whole range of temperature required is only about two 
degrees ; for we know that water boils at 212° under the ordinary 
pressure of the atmosphere in a metallic, and at 214° in a glass 
vessel. In experiments, then, of this nature, the whole quantity 
of mercury which is expelled from the ball, and which rises in the 
stem of the instrument, ere the temperature arrives at 212°, is 
here of no service, and consequently the same length of stem might 
also be dispensed with. 
It is quite evident, that in all experiments such as I here al- 
lude to, where the range of temperature required does not exceed 
a few degrees of Fahrenheit’s scale, the object is to have these 
few degrees extended npon as great a length of scale as possible, 
so as to combine with a convenient length of the instrument 
the nicest delicacy, whereby we may be able to discover the slight- 
est change that takes place in the temperature of the liquids, ei- 
ther by the introduction of foreign substances into them, or by 
varying their circumstances. 
This object I have endeavoured in some degree to attain by the 
following construction, which, by a very simple contrivance, is 
made to contain the portion of mercury which is expelled from the 
ball by expansion, until it arrives at the temperature we wish to 
observe, when it enters a very fine tube, where the most minute 
changes of temperature may be discovered. 
It consists of a ball A, Plate VII. Fig. 1. fully an inch in dia- 
meter, from which rises a short stem CD, about half an inch in 
length and a quarter in diameter ; to its extremity is attached a 
thermometer tube EF of a very fine bore. A portion of mercury 
is then introduced into the ball, so as to fill it, and just rise into the 
tube CD at a natural temperature. This tube CD is formed so as 
to contain the quantity of mercury which expands in the ball be- 
tween the natural temperature and the point, which, in the present 
instance, we are supposing to be 212°. When the liquid has at- 
tained this degree of heat, the mercury begins to ascend in the small 
tube EF, which, on account of the fine bore and the largeness of 
the ball, gives nearly four inches for each degree of Fahrenheit’s 
scale, and thus the slightest variations of temperature may be dis- 
covered. 
It will be seen that the principle may be applied to any tempe- 
rature, all that is necessary is to have the tube CD made so as to 
contain the quantity of mercury that expands between that of the 
natural temperature and the point of investigation. 
It is obvious that a thermometer, where the ball is so large, can- 
not be used to ascertain slight differences of temperature unless 4 
